Synthesis, ligand-receptor modeling studies and pharmacological evaluation of novel 4-modified-2-aryl-1,2,4-triazolo[4,3-a]quinoxalin-1-one derivatives as potent and selective human A3 adenosine receptor antagonists

Article abstract of DOI:10.1016/j.bmc.2008.04.039

The study of some 4-substituted-2-aryl-1,2,4-triazolo[4,3-a]quinoxalin-1-one derivatives, designed as hA₃ adenosine receptor antagonists, is reported. The new compounds bear on the four-position different acylamino, sulfonylamino, benzylureido and benzyloxy moieties, which have also been combined with a para-methoxy group on the 2-phenyl ring or with a nitro residue at the six-position. Many derivatives show high hA₃ adenosine receptor affinities and selectivities both versus hA₁ and hA_2A receptors. The observed structure-affinity relationships of this class of antagonists have been exhaustively rationalized using the recently published ligand-based homology modeling (LBHM) approach. The selected 4-bismethanesulfonylamino-2-phenyl-1,2,4-triazolo[4,3-a]quinoxalin-1-one (13), which shows high hA₃ affinity (K_i = 5.5 nM) and selectivity versus hA₁, hA_2A (both selectivity ratios > 1800) and hA_2B (cAMP assay, IC₅₀ > 10,000 nM) receptors, was tested in an in vitro rat model of cerebral ischemia, proving to be effective in preventing the failure of synaptic activity, induced by oxygen and glucose deprivation in the hippocampus, and in delaying the occurrence of anoxic depolarization.

Full text of DOI:10.1016/j.bmc.2008.04.039

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry 16 (2008) 6086–6102
Synthesis, ligand–receptor modeling studies and pharmacological
evaluation of novel 4-modified-2-aryl-1,2,4-triazolo[4,3-a]
quinoxalin-1-one derivatives as potent and selective
human A3 adenosine receptor antagonists
Vittoria Colotta,a,* Daniela Catarzi,a Flavia Varano,a Ombretta Lenzi,a
Guido Filacchioni,a Claudia Martini,b Letizia Trincavelli,b Osele Ciampi,b Chiara Traini,c
Anna Maria Pugliese,c Felicita Pedata,c Erika Morizzod and Stefano Morod
aDipartimento di Scienze Farmaceutiche, Laboratorio di Progettazione, Sintesi e Studio di Eterocicli Biologicamente Attivi,
Universita’ di Firenze, Polo Scientifico, Via Ugo Schiff, 6, 50019 Sesto Fiorentino, FI, Italy
bDipartimento di Psichiatria, Neurobiologia, Farmacologia e Biotecnologie, Via Bonanno, 6, 56126 Pisa, Italy
c
`
Dipartimento di Farmacologia Preclinica e Clinica, Universita di Firenze, Viale Pieraccini 6, 50139 Firenze, Italy
`
Molecular Modeling Section, Dipartimento di Scienze Farmaceutiche, Universita di Padova, via Marzolo 5, 35131 Padova, Italy
d
Received 9 January 2008; revised 9 April 2008; accepted 18 April 2008
Available online 24 April 2008
Abstract—The study of some 4-substituted-2-aryl-1,2,4-triazolo[4,3-a]quinoxalin-1-one derivatives, designed as hA3 adenosine
receptor antagonists, is reported. The new compounds bear on the four-position different acylamino, sulfonylamino, benzylureido
and benzyloxy moieties, which have also been combined with a para-methoxy group on the 2-phenyl ring or with a nitro residue at
the six-position. Many derivatives show high hA3 adenosine receptor affinities and selectivities both versus hA1 and hA2A receptors.
The observed structure–affinity relationships of this class of antagonists have been exhaustively rationalized using the recently pub-
lished ligand-based homology modeling (LBHM) approach. The selected 4-bismethanesulfonylamino-2-phenyl-1,2,4-triazolo[4,3-
a]quinoxalin-1-one (13), which shows high hA3 affinity (Ki = 5.5 nM) and selectivity versus hA1, hA2A (both selectivity
ratios > 1800) and hA2B (cAMP assay, IC50 > 10,000 nM) receptors, was tested in an in vitro rat model of cerebral ischemia, proving
to be effective in preventing the failure of synaptic activity, induced by oxygen and glucose deprivation in the hippocampus, and in
delaying the occurrence of anoxic depolarization.
ꢀ 2008 Elsevier Ltd. All rights reserved.
1. Introduction
sitive manner (A2A and A2B)1 but it can also modulate
other intracellular signaling pathways. In particular,
the A3 receptor subtype is positively coupled to phos-
pholipase C3 and D,4 KATP channels5 and it was also
proven to induce calcium mobilization.5 The A3 AR is
widely distributed in mammals, and in humans it is
highly expressed in lung and liver, with lower levels in
aorta, brain, and testes.6 The growing understanding
of the physiological effects mediated by the A3 subtype,
such as modulation of cerebral and cardiac ischemia,6,7
inflammation,8 or normal and tumor cell regulation,9
makes this receptor subtype an interesting target for var-
ious therapeutic interventions.10 A3 AR antagonists are
being investigated as potential agents in renal injury,11
neuroprotective agents12,13 and, recently, they have been
described as potential therapeutics in the treatment of
glioblastoma multiforme.14
The modulator adenosine plays important roles in a
wide variety of physiological functions, both in the cen-
tral nervous system and in the periphery. Adenosine ex-
erts its effects through activation of four cell membrane
receptors, termed A1, A2A, A2B, and A3, which belong to
the super family of G protein-coupled receptors.1,2 Acti-
vation of adenosine receptors (ARs) modulates adenylyl
cyclase activity, either in a negative (A1 and A3) or po-
Keywords: G protein-coupled receptors; Adenosine receptor antago-
nists; Triazolo[4,3-a]quinoxalin-1-ones; Tricyclic heteroaromatic sys-
tems; Ligand–receptor modeling studies; Neuroprotective agents.
*
Corresponding author. Tel.: +39 55 5253731; fax: +39 55 5253780;
e-mail: vittoria.colotta@unifi.it
0968-0896/$ - see front matter ꢀ 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2008.04.039

V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
6087
In the past few years, we have directed some of our re-
search toward the study of AR antagonists belonging
to different tricyclic systems.15–23 Among them, the 2-
thesized a new series of 1,2,4 triazolo[4,3-a]quinoxalin-1-
one derivatives (Chart 1 and Table 1), bearing on the 4-
amino group some acyl, sulfonyl, and carbamoyl resi-
dues, which were also combined with a methoxy substi-
tuent or a nitro group at the R1 and R6 positions,
respectively (compounds 1–19). The 4-benzyloxy deriva-
tives 20–21 were also prepared with the purpose of eval-
uating whether the 4-NHCO-group was essential for the
anchoring at the hA3 receptor site. Molecular modeling
studies were carried out on the new derivatives in order
to rationalize their hA3 affinity data and describe their
putative binding mode in our model of the hA3AR rec-
ognition site. Moreover, a selected compound, the 4-bis-
methanesulfonylamino-2-phenyl derivative 13, was
tested in an in vitro rat model of cerebral ischemia to as-
sess its effect in the failure of synaptic activity induced
by oxygen and glucose deprivation.
aryl-1,2,4-triazolo[4,3-a]quinoxalin-1-one
derivatives
were widely investigated by probing different substitu-
ents on the 2-phenyl moiety, the fused benzo ring, and
the four-position (Chart 1). These studies led to the
identification of some structural elements which shift
the affinity toward the human A3 (hA3) receptor sub-
type. The crucial groups which afforded nanomolar
affinity and high selectivity were a 4-methoxy group on
the 2-phenyl ring,16 acyl moieties on the 4-amino
group,16 or a 6-nitro substituent on the fused benzo
ring.17 Interestingly, hindering substituents on the
4-amino group were also well tolerated.22 The contemporary
functionalization of the 2-phenyl-1,2,4-triazolo[4,3-
a]quinoxalin-1-one scaffold with the above cited substit-
uents generally maintained high hA3 AR affinity and
selectivity.22 Molecular modeling studies were carried
out to rationalize the observed structure–affinity rela-
tionships (SAR) and to recognize the hypothetical bind-
ing motif of these compounds in our rhodopsin-based
homology model of human A3 AR. Several hydrogen-
bonding interactions seem to stabilize the anchoring of
the ligands inside the binding cleft.19,20,22 These bonds
involve the 1-carbonyl group, the 6-nitro substituent,
and the 4-acylamino chain, at the level of both NH
and CO functions.20,22 Interestingly, the two hydrogen
bonds of the 4-NH–CO moiety appropriately orient
the R residue inside a hydrophobic pocket. Also the 2-
aryl and the fused benzo ring are located in two size-lim-
ited pockets. Obviously, although most of the previously
investigated triazoloquinoxalin-1-one derivatives share a
similar binding mode inside the transmembrane region
(TM) of the hA3 receptor, each compound finds a fine
adjustment in the receptor cavity to optimize the electro-
static and steric interactions depending on the nature of
the substituents on the triazoloquinoxalin-1-one scaf-
fold. In this context, the substituent on the 4-amino
group seems to play a critical role. On this basis, we syn-
2. Chemistry
The target derivatives 1–21 (Chart 1 and Table 1) were
prepared as depicted in Schemes 1–4.
Synthesis of the 2-aryl-1,2,4-triazolo[4,3-a]quinoxalin-1-
one derivatives 1–3 and 4–6, bearing, respectively, 4-aro-
ylamino moieties or the 4-pyridylcarbonylamino group,
was performed by reacting the suitable 4-amino-triazol-
oquinoxalines 22,16 23,16 or 2417 with the corresponding
acyl chlorides in anhydrous dichloromethane or pyri-
dine (Scheme 1). The 4-sulfonylamino-triazoloquinoxa-
line derivatives 7–12 were obtained allowing the 4-
amino-triazoloquinoxaline derivatives 22–24 to react
with the suitable sulfonyl chlorides (Scheme 2). The 4-
bismethanesulfonylamino-derivatives 13–14 were syn-
thesized by reacting 22 and 24 with an excess of metha-
nesulfonylchloride (Scheme 2). The 4-ureido-
triazoloquinoxaline derivatives 15–19 were obtained
when compounds 22–24 were refluxed with different iso-
cyanates in anhydrous tetrahydrofuran (Scheme 3). The
O
N
R
HN
N
N
N
R1
R6
O
COR
H
SO2Me
SO2R
CONHR
OCH2Ph
N
N
H, NO2
R6
N
N
N
N
N
R1
R1
H, OMe
N
N
H, OMe
O
O
1-19
20-21
Chart 1. Previously and herein-reported 4-modified 2-aryl-1,2,4-triazolo[4,3-a]quinoxalin-1-one derivatives.

6088
V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
O
R
HN
NH2
N
N
N
N
N
N
R1
N
R1
R6
N
R6
a or b
O
O
1-6
22-24
R
R1
H
R6
H
1 C6H4-p-COOMe
R1
22 H
23 OMe
24 H
R6
H
H
2 C6H4-p-COOMe OMe
H
H
H
H
3 C6H4-m-I
4 4-Pyridyl
5 4-Pyridyl
6 4-Pyridyl
H
H
OMe
H
NO2
NO2
Scheme 1. Reagents and conditions: (a) RCOCl, anhydrous pyridine, reflux 10–20 h, 80–97%; (b) RCOCl, anhydrous CH2Cl2 and pyridine, reflux
3 h, 79%.
obenzyl)-50-(N-methylcarbamoyl)adenosine ([125I]AB-
MECA) from cloned hA3 receptor stably expressed in
SO2Ph
HN
R1
H
R6
H
H
N
N
CHO cells. Subsequently, the selected compounds 4–5,
7–8, 13, 15–17, and 20–21, which showed high to good
hA3 AR affinity (Ki < 100 nM), were tested for their
ability to displace [3H]DPCPX from cloned hA1 AR
and [3H]50-(N-ethyl-carboxamido)adenosine ([3H]NECA)
from cloned hA2A ARs, in order to establish their A3
versus A1 and A3 versus A2A selectivity within the same
species. The binding data of 1–21 are shown in Table 1,
together with those of the 4-benzoylamino-1,2-dihydro-
2-phenyl-1,2,4-triazolo[4,3-a]quinoxalin-1-one A,22 theo-
phylline, and DPCPX, reported as comparison. The
4-bismethanesulfonylamino derivative 13, which showed
one of the highest hA3 receptor affinities and selectivities
versus hA1 and hA2A subtypes, among the herein re-
ported compounds, was selected for evaluating: (a) its
potency toward the hA3 receptor by determining the
inhibitory effect on NECA-inhibited cAMP production
in hA3 transfected CHO cells; (b) its affinity and
selectivity toward the hA2B receptor by evaluating its
effect on NECA-stimulated cAMP levels in hA2B trans-
fected CHO cells.
7, 10, 22
N
R1
R6
N
8, 11, 23 OMe
9, 12, 24
O
H
NO2
7-9
a
SO2Me
HN
NH2
N
N
N
N
N
N
R1
N
R1
R6
N
R6
O
O
10-12
22-24
c
SO2Me
N
SO2Me
N
R1
H
N
13, 22
N
R1
N
14, 23 OMe
O
13-14
Scheme 2. Reagents and conditions: (a) PhSO2Cl, anhydrous CH2Cl2
and/or pyridine, reflux 36–140 h, 64–77%; (b) MeSO2Cl, anhydrous
CH2Cl2 and/or pyridine, reflux 14–72 h, 30–50%; (c) excess of
MeSO2Cl, anhydrous CH2Cl2 and pyridine, reflux 15–38 h, 60–85%.
Derivative 13 was also tested to assess its affinity at rat
(r) A1 and A3 ARs. The rA1 affinity was evaluated by
competition experiments using [3H]DPCPX in rat cere-
bral cortex and the rA3 affinity was determined using
[3H](R)-PIA in rat testis membranes, in the presence of
DPCPX to block A1 binding sites. Finally, compound
13 was tested in electrophysiological experiments in rat
hippocampus to evaluate its efficacy in delaying the irre-
versible failure of synaptic activity induced by oxygen
and glucose deprivation.
4-benzyloxy-triazoloquinoxalin-1-one derivatives 20–21
were prepared from the corresponding 4-chloro deriva-
tives 25–2616 and benzyl alcohol, in the presence of so-
dium hydride in dimethylsulfoxide (Scheme 4).
3. Pharmacology
Compounds 1–21 were tested for their ability to displace
[3H]8-cyclopentyl-1,3-dipropylxanthine ([3H]DPCPX)
from A1 AR in bovine cerebral cortical membranes,
[3H]2-[4-(2-carboxyethyl)phenethyl]amino-50-(N-ethyl-
carbamoyl)adenosine ([3H]CGS 21680) from A2A AR in
bovine striatal membranes, and [125I]N6-(4-amino-3-iod-
4. Results and discussion
4.1. Structure–affinity relationships
The binding results reported in Table 1 indicate that
some of the newly synthesized 1,2,4-triazolo[4,3-a]qui-

V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
6089
O
R
N
H
NH2
N
HN
N
N
N
N
a
N
R1
N
R1
R6
R6
N
O
O
22-24
15-19
R
R1
H
R6
H
H
NO2
H
R1
22 H
23 OMe
24 H
R6
H
H
15 CH2Ph
16 CH2Ph OMe
17 CH2Ph
18 COPh
19 Ph-m-I
H
H
H
NO2
H
Scheme 3. Reagents and conditions: (a) R–N@C@O, anhydrous tetrahydrofuran, reflux 3–36 h, 55–95%.
Ph
O
Cl
N
N
N
N
N
R1
a
N
R1
+
Ph
OH
N
N
O
O
20-21
25-26
R1
H
25, 20
26, 21 OMe
Scheme 4. Reagents and conditions: (a) 60% NaH dispersion in mineral oil, anhydrous dimethylsulfoxide, rt 12–15 h, 35–48%.
noxalin-1-ones show high hA3 receptor affinity and
selectivity (compounds 7–8, 13, 16, 17, and 20–21) both
versus hA1 and hA2A ARs.
compound 7 with a 4-methanesulfonylamino chain
(compound 10) dramatically decreased the hA3 AR
affinity (Ki = 1427 nM), probably because of the reduc-
tion of the lipophilicity of the substituent and, conse-
quently, of the weaker hydrophobic interaction of
compound 10 with the receptor site. When a second
methanesulfonyl group was introduced on the 4-amino
moiety of 10, the hA3 affinity significantly ameliorated.
Indeed, compound 13 binds the hA3 subtype with nano-
molar affinity (Ki = 5.5 nM) and also with high selectiv-
ity, versus both hA1 and hA2A receptors. To determine
also its hA3 versus hA2B selectivity, we tested compound
13 in cAMP assays, which evidenced a lack of hA2B
affinity, being 13 ineffective in inhibiting NECA-stimu-
lated cAMP levels in hA2B CHO cells (IC50 > 10,000).
Considering the previously reported 4-benzoylamino-
1,2-dihydro-2-phenyl-1,2,4-triazolo[4,3-a]quinoxalin-1-
one A22 as reference compound, we can observe that
introduction of a para-carbomethoxy group or a meta-
iodo substituent on the benzoyl moiety was deleterious
for the hA3 receptor affinity (compounds 1 and 3), prob-
ably because of the steric bulk enhancement of the 4-
aroylamino portion. The same negative effect was pro-
duced on affinities at all the other investigated receptors.
Isosteric replacement of the benzoyl residue of A with a
4-pyridylcarbonyl moiety (compound 4) maintained the
hA3 affinity in the low nanomolar range and decreased
the capability to bind the hA1 receptor, thus enhancing
the hA3 versus hA1 selectivity. Compound 4 also dis-
plays some (30-fold) hA3 versus hA2A receptor selectiv-
ity. Replacement of the 4-benzoylamino group of A with
the 4-benzenesulfonylamino moiety (compound 7) was
made because we hypothesized that the sulfonyl group
could engage a hydrogen bond with the receptor site,
similarly to the benzoyl carbonyl group of compound
A. This structural modification, although affording some
reduction of the hA3 AR affinity (Ki = 32 nM), signifi-
cantly increased the selectivity for the hA3 receptor sub-
type, being both hA1 and hA2A receptor affinities null.
Replacement of the 4-benzenesulfonylamino moiety of
Among the probed ureido substituents (compound 15–
19) only the benzylureido chain afforded a significant
hA3 affinity (compounds 15–17). A very interesting re-
sult was found for the 4-benzyloxy derivative 20, which
possesses a high affinity for the hA3 AR (Ki = 21 nM),
although lacking the 4-NH–CO– group. This derivative
is also highly hA3 selective, in comparison to both the
hA1 and hA2A
.
The para-methoxy substituent at the R1 position was
beneficial in compounds 2, 8, 11, and 21 since they show
higher hA3 AR affinities with respect to the correspond-
ing 2-phenyl derivatives 1, 7, 10, and 20. In the other

6090
V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
Table 1. Binding affinity at human A1, A2A, A3 and bovine A1, and A2A ARs
R4
N
N
N
R1
R
N
6
O
A, 1-21
Compound
R4
R1
R6
Kia (nM) or I%
b
c
d
e
f
hA3
hA1
hA2A
bA1
bA2A
Ag
NHCOPh
H
H
1.47 0.06
41%
87.8 6.3
88.2 6.7
89.6 7.2
106 2.1
30.5%
53%
36%
41%
35%
1
NHCOC6H4-4COOMe
NHCOC6H4-4COOMe
NHCOC6H4-3I
NHCO-4-Pyridyl
NHCO-4-Pyridyl
NHCO-4-Pyridyl
NHSO2Ph
H
H
2
OMe
H
H
1370 121
36%
3
H
473 34
57 4.3
236 15
37.5%
4
H
H
6.1 0.5
68 5.2
0%
2379 191
779 53
188 9.4
397 39
812 71
44%
22%
5
OMe
H
H
6
NO2
H
7
H
32.2 2.8
2.2 0.11
100 7.2
1427 125
493 33
37%
0%
2700 142
27%
23%
157 1.4
4700 260
210 12
164 11.3
6%
35%
16%
8
NHSO2Ph
NHSO2Ph
OMe
H
H
9
10
NO2
H
25%
32%
NHSO2CH3
NHSO2CH3
H
11
12
OMe
H
H
0%
25%
NHSO2CH3
N(SO2CH3)2
N(SO2CH3)2
NO2
H
2840 162
36 1.3
6.2%
13
14
H
5.5 0.4
387 24
83.5 4.9
65 5.1
63 4.4
1300 115
953 61
21 1.8
6.4 0.4
86,000 7800
1300 125
36%
32%
56%
17%
OMe
H
H
15
16
NHCONHCH2Ph
NHCONHCH2Ph
NHCONHCH2Ph
NHCONHCOPh
NHCONH-Ph-3I
OCH2Ph
H
12.3 1.2
4215 350
4%
158.3 15
23%
20%
4.1 0.2
20.8 1.2
4.6 0.3
100.6 8.9
359 25
55 3.6
53 %
172.6 12
12%
OMe
H
H
17
18
NO2
H
46.5%
379 24
1800 150
19%
H
19
20
H
H
H
OMe
H
46%
54%
10%
4%
21
Theophylline
DPCPX
OCH2Ph
H
41%
21,000 1800
337 28
6200 530
3.2 0.2
7900 630
260 18
3800 340
0.5 0.03
a The Ki values are means SEM of four separate assays, each performed in triplicate.
b Displacement of specific [125I]AB-MECA binding at hA3 receptors expressed in CHO cells or percentage of inhibition (I%) of specific binding at
1 lM concentration.
c Displacement of specific [3H]DPCPX binding at hA1 receptors expressed in CHO cells or percentage of inhibition (I%) of specific binding at 10 lM
concentration.
d Displacement of specific [3H]NECA binding at hA2A receptors expressed in CHO cells or percentage of inhibition (I%) of specific binding at 10 lM
concentration.
e Displacement of specific [3H]DPCPX binding in bovine brain membranes or percentage of inhibition (I%) of specific binding at 10 lM
concentration.
f Displacement of specific [3H]CGS 21680 binding from bovine striatal membranes or percentage of inhibition (I%) of specific binding at 10 lM
concentration.
g Ref. 22.
cases, the para-methoxy group left almost unchanged
(compare compound 16 with 15) or worsened (compare
compounds 5 and 14 with 4 and 13, respectively) the
binding at the hA3 receptor.
in inhibiting the NECA-inhibited cAMP accumulation
in hA3 CHO cells showing an antagonistic behaviour
(Fig. 1, EC50 = 12 1.1 nM). Since compound 13 was
tested in an in vitro rat model of cerebral ischemia, we
also evaluated its rA1 and rA3 AR affinities. As ex-
pected, the rat A3 affinity was very low. At 1 lM con-
centration, it inhibited the radioligand-binding by only
5%. On the contrary, the ability of compound 13 to dis-
place rA1 ligand from its receptor was significantly high-
er (Ki = 132 12 nM) than shown for hA1 receptor.
Introduction of the nitro group at the R6 position was
not beneficial as it reduced the hA3 affinity (compare 7
with 9 and 10 with 12) or annulled it (compare 4 with
6). The only exception is represented by the 4-benzylure-
ido derivative 17, which possesses the same good hA3
AR affinity of the 6-unsubstituted parent derivative 15,
and also higher hA3 receptor selectivity.
The bA1 and bA2A binding data of the new derivatives
1–21 (Table 1) warrant some comments. As expected,
none of the compounds significantly bind to the bA2A
AR, with the exception of three derivatives (4, 15, and
18). In contrast, several derivatives (1, 4, 13, 15, and
18–20), all lacking both the R1 and R6 substituents, pos-
The selected 4-bismethanesulfonylamino derivative 13
was tested to assess its potency versus the hA3 receptor.
Consistently with its high hA3 affinity value
(Ki = 5.5 nM), compound 13 proved to be highly potent

V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
6091
the hA1 one, in perfect accordance with the results ob-
tained in previous studies, either on this19,22 or other
classes of derivatives.20,21,23–26 These results confirm
that, notwithstanding the high sequence homology
(94%) between the human and bovine A1 receptors,24
significant structural differences exist in the region of
the recognition site of the two receptors.
60
50
40
30
20
10
0
4.2. Molecular modeling studies
-11
-10
-9
-8
-7
-6
Molecular modeling studies were performed on the 2-
aryl-1,2,4-triazolo[4,3-a]quinoxalin-1-one derivatives A
and 1–21 in order to identify the hypothetical binding
motif of the new hA3 antagonists and rationalize the ob-
served SAR.
log [compound]
Figure 1. Effect of compound 13 on NECA-mediated cAMP accumu-
lation (vs forskolin, set to 100%) in CHO cells stably expressing hA3
AR. Data represents means SEM from three separate experiments.
The IC50 value for 13 was 12 1.2 nM.
Following our previously reported modeling studies,19–
23
we have constructed a refined model of hA3 receptor
sess high to good bA1 affinities. As previously observed
in this class of AR antagonists,16,19,22 the methoxy
group at the R1 position decreases the binding at the
bA1 receptor. Indeed, all the 2-(4-methoxyphenyl)-deriv-
atives (2, 5, 8, 11, 14, 16, and 21) show lower bA1 recep-
tor affinities than those of the corresponding 2-phenyl-
substituted compounds (1, 4, 7, 10, 13, 15, and 20).
Instead, the nitro group at the R6 position does not exert
a constant effect: derivatives 6 and 12 show significantly
reduced bA1 affinities, in comparison to their 6-des nitro
compounds 4 and 10, while compounds 9 and 17 display
similar bA1 AR affinities to those of derivatives 7
and 15.
by using a rhodopsin-based homology modeling
(RBHM) approach.27–30 Moreover, our recently de-
scribed ligand-based homology modeling (LBHM) ap-
proach has been used to simulate the conformational
changes induced by ligand binding, and a schematic rep-
resentation of the ligand-based homology protocol is
shown in Figure 2 (methodological details are summa-
rized in Section 6.2).31 As reported in Figure 3, depend-
ing on the topological properties of the different ligands,
we found four different conformational models of the
human A3 receptor reverse agonist-like state in which
both shape and chemical complementarities have been
specifically optimized around each ligand. In this spe-
cific case, with varying ligand structure, the molecular
volume of the transmembrane (TM) binding cavity
It is worth noting that these compounds show signifi-
cantly different affinities at the bA1 receptor than at
3
˚
changes from the 660 A of the standard RBHM-driven
Figure 2. Flow chart of the ligand-based homology modeling (LBHM) technique considering an evolution of a conventional rhodopsin-based
homology modeling (RBHM) algorithm.

6092
V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
Figure 3. Ligand-based homology modeling (LBHM) data collection. Each ‘reference derivative’ (compounds A, 13 and 15) was used as ligand-
template during the homology modeling process. Consequently, four different conformational states (models 1–4) were selected as putative
ambassadors of the conformational changes induced by different ligand binding. Depending on their different structure topologies, all other
antagonists (docked derivatives) were docked into the most complementary receptor models.
3
˚
model to the 1120 A of the largest LBHM-driven mod-
el, without altering the conventional rhodopsin-like
receptor topology. The modifications of both shape
and volume of the human A3 TM binding cavity are
the most important receptor modeling perturbations ob-
tained by the application of the LBHM technique. The
binding cavity reorganization induced by ligand binding
is due to the conformational change in several amino
acid side chains, such as Leu90 (3.32), Leu91 (3.33),
Thr94 (3.36), His95 (3.37), Ile98 (3.40), Gln167 (EL2),
Phe168 (EL2), Phe182 (5.43), Ile186 (5.47), Leu190
(5.51), Phe239 (6.44), Trp243 (6.48), Leu244 (6.49),
Leu264 (7.35), and Ile268 (7.39). However, molecular
docking studies carried out for all the triazoloquinoxa-
line antagonists, using the appropriate conformational
states of the receptor as listed in Figure 3, have shown
a similar binding motif indicating that a common recep-
tor-driven pharmacophore model can be depicted. This
finding is coherent with our previously reported stud-
ies.23,27–31 Interestingly, none of the new triazoloquinox-
aline antagonists found an energetically stable docking
pose in the conventional RBHM-driven A3 model. This
is mainly due to the unfavorable topological comple-
mentarity among these antagonists and corresponding
RBHM-driven TM binding cavity. In particular, highly
destabilizing van der Waals interactions (steric conflicts)
seem to be the reason for lacking topological comple-
mentarities. These steric conflicts are drastically reduced
or completely eliminated after the application of the
LBHM approach.
quinoxaline moiety is surrounded by TMs 3, 5, 6, and 7
with the substituent in the four-position oriented toward
the intracellular environment. Furthermore, this hypo-
thetical binding cleft has also been recently suggested
by other authors.32,33 As shown in Figure 4, the phenyl
ring at the two-position is close to TMs 3, 6, and 7. Ana-
lyzing our model in detail, all triazoloquinoxaline deriv-
atives share at least two stabilizing hydrogen-bonding
interactions inside the binding cleft. The first hydrogen
bond is between the carbonyl group at one-position,
that points toward the EL2, and the NH2 of the
Gln167. This hydrogen-bonding distance is calculated
around 2.8 A for all docked compounds. Moreover,
the 1-carbonyl group is also at the hydrogen-bonding
˚
distance (ca. 3.2 A) with the amide moiety of Asn250
(6.55) side chain. This asparagine residue, conserved
among all adenosine receptor subtypes, was found to
be important for ligand binding.34,35
˚
An important hydrogen-bonding network can be ob-
served in all energetically stable docked conformations
of all the triazoloquinoxaline antagonists; in particular,
Thr94 (3.36), His95 (3.37), and Ser247 (6.52) are able to
interact through hydrogen bonds with the 4-carbonyl
oxygen of compounds 1–19 and with the ether oxygen
of derivatives 20–21. These polar amino acids seem to
be critical for the recognition of all antagonist structures
and for receptor selectivity. In particular, Ser247 (6.52)
of the hA3 receptor subtype is not present in the corre-
sponding position of hA1 and hA2 receptors, where
the residue is replaced by a histidine (His251 in hA1,
His250 in hA2A, and His251 in hA2B). The histidine side
chain is bulkier than serine and, possibly for this reason,
As previously described,19–22 ligand recognition occurs
in the upper region of the TM bundle, and the triazolo-

V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
6093
Figure 4. Hypothetical binding motif of the representative newly synthesized triazoloquinoxaline antagonists. The most energetically favorable
docked conformation of derivative 4 into LBHM-model 1 is viewed from the membrane side facing TM helices 4 and 5. Side chains of some amino
acids important for ligand recognition are highlighted. Hydrogen atoms are not displayed. Moreover, the receptor region around R4-substituents
characterized by five non-polar amino acids, Ile98 (TM3), Ile186 (TM5), Phe239 (TM6), Phe243 (TM5), and Ser271 (TM7), has been represented by
its Connolly’s molecular surface.
large substituents at the four-position of the triazoloqui-
noxaline framework are not well tolerated by hA1 and
hA2 receptor subtypes. Indeed, 4-acylamino, 4-sulfon-
amido and 4-benzylureido derivatives are inactive or
modestly active on hA1 and hA2A ARs. On the contrary,
the hydroxyl group of Ser247 (6.52) of the hA3 receptor
is appropriately positioned to form a hydrogen-bonding
interaction with the carbonyl oxygen of the 4-amido/sul-
fonamido/ureido group of compounds A, 1–19. In par-
ticular, the 4-sulfonamido derivatives 13 and 14
interact simultaneously through hydrogen bonds with
all three polar amino acids Thr94 (3.36), His95 (3.37),
and Ser247 (6.52). Interestingly, also the 4-benzyloxy
analogs 20–21 are selectively accommodated into the
hA3 binding cavity. These observations support the
importance of the group at the four-position in modu-
lating receptor selectivity. Indeed, the receptor region
around the R4-substituent is mostly hydrophobic and
characterized by five non-polar amino acids: Ile98
(3.40), Ile186 (5.47), Leu190 (5.51), Phe239 (6.44), and
Leu244 (6.49), as shown in Figure 4.
valine (Val87) in the human A1 receptor. This mutation
might explain the hA3 versus hA1 selectivity. In fact,
even if the mutation Leu90 (hA3)/Val87 (hA1) can
slightly enlarge the dimension of this hydrophobic cav-
ity, simultaneously it also sensibly decreases both shape
and hydrophobic complementarities (data not shown).
Also the mutation of Ser165 (EL2 of hA3) with
Lys168 in the hA1 receptor could affect the recognition
of the
derivatives.
methoxy-substituted
triazoloquinoxaline
As previously described,22 the presence of the 6-nitro
substituent has not always produced advantageous ef-
fects in terms of hA3 AR binding affinity. This phe-
nomenon is particularly evident comparing derivatives
6 and 12 with respect to their unsubstituted com-
pounds 4 and 10. As already anticipated and clearly
shown in Figure 3, the relative position of R6-substitu-
ent is slightly different depending on the bulkiness of
the R4-substituent on the carbamoyl moiety at the
four-position. In particular, in the presence of a less
bulky R4 substituent, the triazoloquinoxaline moiety
binds more deeply in the middle of the TM bundle,
positioning the 6-nitro substituent very close to TM5.
In this case, unfavorable steric and dipolar interactions
are responsible for the remarkable reduction of affinity
observed for derivatives 6 and 12. Increasing the bulk-
iness of the R4-substituent, the position of the R6
group shifts away from TM5 and, consequently, more
empty space is available for the 6-nitro substituent such
as in derivatives 9 and 17.
Considering the observed structure–activity relation-
ships in greater detail, methoxy substitution at R1 posi-
tion is rather well tolerated among all newly synthesized
triazoloquinoxaline derivatives. This is consistent with
its accommodation into a tiny hydrophobic pocket
delimited by Leu90 (3.32) and Ile268 (7.39). Interest-
ingly, the amino acid corresponding to Leu90 in the
hA3 receptor was found to be essential for the binding
of both agonists and antagonists, and it is mutated in

6094
V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
4.3. Electrophysiological studies
A
compound 13
7 min OGD
7 min OGD
Compound 13 was also tested in an in vitro rat model of
cerebral ischemia obtained by oxygen and glucose depri-
vation (OGD). During cerebral ischemia, extracellular
adenosine concentration drastically increases both
in vivo and in vitro36 reaching a level able to stimulate
all adenosine receptor subtypes, A1, A2A, A2B, and A3.
Recently, it has been demonstrated that, in the CA1 hip-
pocampus in vitro, the role played by A3 receptors under
ischemia depends upon the time duration of ischemic in-
sults.37 A3 receptors stimulated by adenosine released
during brief periods of ischemia, obtained by OGD,
might exert A1-like protective effects on neurotransmis-
sion. On the contrary, prolonged periods of OGD would
transform the A3 receptor-mediated effects from protec-
tive to injurious.37 The latter result is in agreement with
that obtained in the same brain area13,23 using different
hA3 AR and rA3AR antagonists.
0
-4
0
-4
-8
-8
-12
-12
0
10
20
0
10
20
Time (min)
Time (min)
B
10 nM compound 13
7 min OGD
110
90
An important factor in the study of brain ischemia is the
generation of anoxic depolarization (AD), a consistent
drop of neuron and glia membrane potential which
may contribute to the amount and severity of neuronal
damage38,39 in this pathological condition. It has been
suggested that a possible strategy in the development
of neuroprotective drugs may include agents that inhibit
the initiation or propagation of AD.
70
50
OGD untreated slices (n=8)
OGD + 10 nM compound 13 (n=4)
30
10
-10
In the present study we investigated the effects of the
newly synthesized compound 13 on synaptic transmis-
sion and AD development under OGD in the CA1 re-
gion of rat hippocampal slices. We examined the
effects of 7 and 30 min OGD in the absence or presence
of derivative 13.
-10
0
10
Time (min)
20
30
Figure 5. The selective A3 antagonist 13 prevents the loss of
neurotransmission induced by 7-min OGD. (A) AD was recorded as
the negative d.c. shift in response to 7-min OGD (solid bar) in the
absence (n = 8, left panel) and in the presence of 10 nM compound 13
(open bar, n = 4, right panel). Note that the A3 antagonist completely
prevented AD appearance. (B) Graphs show the time course of 7-min
OGD (solid bar) effect on fEPSP amplitude, expressed as percentage of
baseline, in untreated OGD slices (filled circles, means SEM, n = 8)
or in the presence of 10 nM compound 13 (open circles, means SEM,
n = 4). Solid bars indicate the duration of OGD.
As shown in Figure 5A (left panel), 7-min OGD con-
stantly elicited AD, recorded as negative d.c. (direct cur-
rent) shifts, with a mean peak latency of 6.3 0.3 min
from the beginning of ischemia and a peak amplitude
of ꢀ7.7 0.8 mV (n = 8). In the presence of compound
13 (10 nM, n = 4) AD was absent in all recorded slices
(Fig. 5A, right panel). In addition, 7-min OGD elicited
a complete loss of fEPSP amplitude which was pre-
vented when OGD was carried out in the presence of
compound 13 (Fig. 5B). The A3 AR antagonist 13 was
applied 15 min before, during, and 5 min after OGD.
The compound did not change fEPSP amplitude under
adenosine. This effect is detectable by first 2 min of
ischemia.40 Nevertheless, at the concentration used in
our experiments, the A1 antagonistic effect of compound
13 (Ki = 132 12 nM) did not appear since the com-
pound did not modify the depression of fEPSP induced
by OGD (Fig. 5B).
normoxic
conditions
(from
1.0 0.04 mV
to
0.97 0.05 mV after 15 min application, n = 4), but it
induced a significant fEPSP recovery after OGD
(100 2%, n = 4, in comparison to 4.2 1.6%, n = 8
found in untreated hippocampal slices). These findings
are in agreement with our previous results showing that
several, chemically diverse, A3 receptor antagonists pre-
vented the irreversible failure of neurotransmission in-
duced by 7-min OGD and induced a complete AD
abolishment or delay.13,25,37
In order to well-define the effects of compound 13 on
AD development, we prolonged the duration of ische-
mic insult from 7 to 30 min, a time duration able to in-
duce an irreversible loss of neurotransmission in the
CA1 hippocampus in vitro.13 In untreated slices, 30-
min OGD always elicited AD with a mean peak time
of 6.9 0.3 min (n = 11), measured from the beginning
of ischemic insult, and an amplitude of ꢀ7.6 0.8 mV
(Fig. 6A). Compound 13 (10 nM, n = 11), applied
15 min before and during OGD, induced a significant
delay in AD appearance (9.3 0.5 min, P = 0.0006,
Fig. 6A and B), without producing any considerable
change in AD amplitude (ꢀ6.3 0.7 mV). Compound
It has to be noted that the most likely mechanism of
fEPSP reduction in the CA1 hippocampus during ische-
mia is a decrease in glutamate release caused by activa-
tion of presynaptic A1 receptors induced by extracellular

V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
6095
sult would appear to be in discrepancy with the very low
rA3 AR binding affinity of 13. However, it is consistent
with previous findings showing that in the same in vitro
model, other A3 antagonists,13,26,37 such as (2R,3R,4S)-
2-(2-chloro-6-(3-iodobenzylamino)-9H-purin-9-yl)tetra-
hydrothiophene-3,4-diol (LJ 1251),43 3-propyl 6-ethyl-5-
[(ethylthio)carbonyl]-2-phenyl-4-propyl-3-pyridinecar-
boxylate (MRS 1523),44 N-[9-chloro-2-(2-furanyl)-1,2,4-
triazolo[1,5-c]quinazolin-5yl]benzeneacetamide (MRS
1220)45 and 5-[[(4-pyridyl)amino]carbonyl]amino-8-
methyl-2-(2-furyl)-pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]-
pyrimidine hydrochloride,46 and 4-amido-2-phenylpy-
razolo[3,4-c]quinoline derivatives23 led to recovery of
synaptic activity after 7-min OGD at nanomolar con-
centration. Among them, only MRS 1523 and LJ 1251
possessed nanomolar affinity for the rA3 AR (Ki = 113
and 3.89 nM, respectively).43,44 As already discussed,37
the apparent discrepancy between the rA3 binding affin-
ity of compound 13 and its protective effect on OGD,
observed at 10 nM concentration, might depend on the
specific environment and coupling of A3 receptors in
the cell membrane of native tissue and may differ from
that assessed in receptor binding experiments. Alterna-
tively, the paucity of A3 AR in native tissue,47,48 allows
one to speculate that occupancy of a substantial fraction
of A3 receptors is required to evoke cell response(s).
Thus, the blockage of a relatively small fraction of A3
receptors may be sufficient to antagonize the effect of
endogenous adenosine released during OGD.
10 nM compound 13
A
30 min OGD
0
-4
-8
OGD untreated slices
OGD+10 nM compound 13
-12
0
10
20
30
Time (min)
P=0.0006
B
10
*
5
(n=11)
(n=11)
0
OGD untreated slices
OGD+compound 13
Figure 6. The selective A3 antagonist 13 delays the appearance of AD
elicited by 30-min OGD in the CA1 region of rat hippocampal slices.
(A). Traces are recordings of negative d.c. shifts in response to 30-min
OGD taken from hippocampal slices in the absence (n = 11, contin-
uous lines) or in the presence of 13 at 10 nM concentration (open bar,
n = 11, dotted lines). Note that in the presence of the selective A3
adenosine receptor antagonist the appearance of AD was always
delayed in comparison to that obtained in the absence of the drug. (B)
Columns represent means SEM of AD peak time recorded in
hippocampal slices during 30-min OGD application, measured from
the beginning of ischemic insult. Compound 13 significantly postponed
AD development. *P = 0.0006, unpaired two-tailed Student’s t test. (n)
indicates the number of tested slices.
There is growing evidence that during ischemia the
development of AD can be an important determinant
of the degree and extent of ischemic damage and that
NMDA receptors are essential to AD initiation and
propagation.39 The time-window of A3 receptor-medi-
ated effects found in the present work, and also reported
by Pugliese et al.,13,37 overlaps with the delay that can be
obtained by treating the slices with glutamate receptor
antagonists.49,50 On this basis, we suppose that the post-
ponement of AD appearance caused by derivative 13
may be attributable to reduction in excitotoxic gluta-
mate effect. Rat cortical neurons exposed to hypoxia
in vitro show an increase in activation of protein kinase
C (PKC) after selective A3 receptor stimulation.51 It has
been demonstrated that in the CA1 region of rat hippo-
campal slices the activation of A3 adenosine receptors
reduces the inhibitory function of metabotropic gluta-
mate receptors on glutamate excitotoxic effect.52 Over-
all, these mechanisms may contribute to a deleterious
role of prolonged stimulation of A3 receptors during
ischemia.
13 was ineffective on AD appearance at a concentration
of 0.5 nM (n = 2, AD peak time of ꢁ7 min, data not
shown), while a significant effect started from 1 nM
(n = 3 AD peak time of 8.7 0.5, data not shown). No
increase in the effect of compound 13 was observed up
to the concentration of 1 lM (n = 3, AD peak time of
9.2 0.5 min, data not shown). After 30-min OGD car-
ried out in the presence of derivative 13, no recovery of
fEPSP amplitude was observed also after prolonged
(40 min n = 3, data not shown) reperfusion in normal
oxygenated aCSF (artificial cerebral spinal fluid).
5. Conclusion
During OGD conditions adenosine is released from hip-
pocampal slices, reaching at the receptors concentra-
tions up to 30 lM after 5-min OGD.40,41 These
concentrations can stimulate A3 ARs which have affinity
in the micromolar range for adenosine.42 We demon-
strated that compound 13 permits the recovery of synap-
tic activity after an otherwise irreversible ischemic insult
(7-min duration) and delays the occurrence of AD
brought about by a prolonged period of OGD. This re-
The present study afforded new potent hA3 AR antago-
nists (compounds 7–8, 13, 16, 17, 20–21), belonging to
the class of the 2-aryl-1,2,4-triazolo[4,3-a]quinoxaline-
1-ones, which also showed high selectivities versus both
hA1 and hA2A ARs.
Molecular docking studies, carried out using a ligand-
based homology modeling approach, led to the identifi-

6096
V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
cation of the putative binding mode of the new deriva-
tives which was coherent with the observed SAR. Com-
pound 13, tested in an in vitro rat model of cerebral
ischemia, prevented the irreversible failure of the neuro-
transmission caused by a 7-min ischemic insult and de-
layed the occurrence of AD induced by a prolonged
(30-min) period of OGD. These results show that deriv-
ative 13 might be efficacious in ameliorating the resis-
tance of the CA1 region of hippocampus to ischemic
damage.
(1.3 mmol) in anhydrous pyridine (6 mL). The mixture
was refluxed for about 10–15 h. After cooling at room
temperature, the mixture was diluted with water
(30 mL) and the solid collected by filtration.
1
Compound 1: Yield: 83%; mp 251–252 ꢁC (DMF). H
NMR 3.93 (s, 3H, OMe), 7.38 (t, 1H, ar, J = 7.4 Hz),
7.54–7.68 (m, 4H, ar), 7.79 (d, 1H, ar), 8.02 (d, 2H, ar,
J = 8.4 Hz), 8.13–8.20 (m, 4H, ar), 8.79 (d, 1H, H-9,
J = 8.1 Hz), 11.48 (br s, 1H, NH). IR 1727.
1
Compound 2: Yield: 90%; mp 256–258 ꢁC (DMF). H
6. Experimental
6.1. Chemistry
NMR 3.09 (s, 3H, OMe), 3.79 (s, 3H, COOMe), 7.09
(d, 2H, ar, J = 7.0 Hz), 7.52–7.89 (m, 5H, ar), 8.08–
8.18 (m, 4H, ar), 8.75 (d, 1H, H-9, J = 8.1 Hz), 11.51
(br s, 1H, NH). IR 1713.
Silica gel plates (Merck F254) and silica gel 60 (Merck,
70–230 mesh) were used for analytical and column chro-
matography, respectively. All melting points were deter-
mined on a Gallenkamp melting point apparatus.
Microanalyses were performed with a Perkin-Elmer
260 elemental analyzer for C, H, N, and the results (re-
ported in Table 2) were within 0.4% of the theoretical
values, unless otherwise stated. The IR spectra were re-
corded with a Perkin-Elmer Spectrum RX I spectrome-
6.1.2.
4-(3-Iodobenzoylamino)-1,2-dihydro-2-phenyl-
1,2,4-triazolo[4,3-a]quinoxalin-1-one (3). A solution of
3-iodobenzoyl chloride54 (3.38 mmol) in anhydrous
dichloromethane (5 mL) was added to a suspension of
2216 (0.68 mmol) in anhydrous dichloromethane
(20 mL) and anhydrous pyridine (9.8 mmol, 0.7 mL).
The mixture was refluxed for 3 h. Evaporation of the
solvent at reduced pressure gave a residue which was
treated with water/ethanol, collected by filtration and
washed with water. Yield: 79%; mp 286–287 ꢁC
1
ter in Nujol mulls and are expressed in cmꢀ1. The H
NMR spectra were obtained with a Brucker Avance
400 MHz instrument. The chemical shifts are reported
in d (ppm) and are relative to the central peak of the sol-
vent which was DMSO-d6. The following abbreviations
are used: s = singlet, d = doublet, dd = double doublet,
t = triplet, m = multiplet, br = broad, and ar = aromatic
protons.
1
(AcOH/DMF). H NMR 7.38–7.98 (m, 6 H, ar), 7.79
(d, 1H, J = 8.1 Hz), 7.98–8.02 (m, 4H, ar), 8.37 (s, 1H,
ar), 8.76 (d, 1H, H-9, J = 8.1 Hz), 11.35 (br s, 1H,
NH). IR 1680, 1735, 3225.
6.1.3. General procedure for the synthesis of 2-aryl-4-(4-
pyridylcarbonylamino)-1,2-dihydro-1,2,4-triazolo[4,3-a]-
quinoxalin-1-ones (4–6). A solution of 4-pyridylcarbon-
ylchloride55 (2.6 mmol) in anhydrous pyridine (2 mL)
was added dropwise at room temperature to a suspen-
sion of compounds 22–2416,17 (1.3 mmol) in anhydrous
pyridine (10 mL). The mixture was refluxed for about
6.1.1. General procedure for the synthesis of 2-aryl-4-(4-
carbomethoxybenzoylamino)-1,2-dihydro-1,2,4-triazolo-
[4,3-a]quinoxalin-1-ones (1–2). A solution of 4-carbo-
methoxybenzoyl chloride53 (2.6 mmol) in anhydrous
pyridine (2 mL) was added to a suspension of 22–2316
Table 2. Combustion Analysis data of the newly synthesized compounds
Compound
Formula
C
Calcd–found
H
Calcd–found
N
Calcd–found
1
2
C24H17N5O4
C25H19N5O5
65.60–65.74
63.96–63.79
52.09–51.05
65.96–66.15
64.07–64.28
59.02–58.94
60.42–60.20
59.05–58.94
54.54–54.78
54.08–54.20
52.98–53.12
48.00–48.23
47.10–47.32
46.64–46.39
67.31–67.23
65.45–65.27
60.66–60.40
50.59–50.38
71.73–71.58
69.34–69.66
3.90–3.78
4.08–4.23
2.78–2.59
3.69–3.88
3.91–3.75
3.07–3.20
3.62–3.82
3.83–3.60
3.05–3.19
3.69–3.88
3.92–3.74
3.02–3.25
3.49–3.60
3.70–3.91
4.42–4.65
4.58–4.81
3.76–3.90
2.89–2.70
4.38–4.61
4.55–4.70
15.94–16.10
14.92–14.75
13.81–14.01
21.98–22.10
20.38–20.10
22.94–22.78
16.78–16.99
15.65–15.44
18.17–18.24
19.71–19.54
18.17–18.01
20.99–20.78
16.16–16.25
15.11–14.91
20.48–20.30
19.08–19.29
21.53–21.42
16.09–16.31
15.21–15.39
14.06–14.29
3
C
C
22H14I N5O2
21H14N6O2
4
5
C22H16N6O3
C21H13N7O4
6
7
C21H15N5O3S
C22H17N5O4S
C21H14N6O5S
8
9
10
11
12
13
14
15
16
17
19
20
21
C16H13N5O3S
C17H15N5O4S
C16H12N6O5S
C17H15N5O5S2
C18H17N5O6S2
C23H18N6O2
C24H20N6O3
C
C
C
23H17N7O4
22H15IN6O2
22H16N4O2
C23H18N4O3

V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
6097
20 h. After cooling at room temperature the solution
was diluted with water (30 mL) and the solid precipitate
was filtered.
additions of the same quantity of methanesulfonyl chlo-
ride (5 mmol, 0.4 mL) were done every 12 h, for a total
reflux time of about 90 h. After cooling, the mixture was
taken up with water (20 mL) and the resulting solid was
collected by filtration and washed with water. The crude
solid was purified by column chromatography, (SiO2,
eluting system cyclohexane/ethyl acetate/methanol,
4.5:4.5:0.5). Evaporation of the second eluates gave a so-
lid which was recrystallized from 2-methoxyethanol.
Yield: 30%; mp 270–272 ꢁC. 1H NMR 3.33 (s, 3H,
SO2CH3), 7.37–7.43 (m, 3H, ar), 7.56–7.60 (m, 2H,
ar), 7.87–8.03 (m, 3H, ar), 8.63–8.67 (m, 1H, H-9),
11.50 (br s, 1H, NH). IR 1644, 1734, 3181.
1
Compound 4: Yield: 85%; mp 273–275 ꢁC (DMF). H
NMR 7.25–7.37 (m, 1H, ar), 7.42–7.82 (m, 5H, ar),
7.85–8.02 (m, 4H, 2-pyridine protons + 2 ar), 8.73–8.82
(m, 3H, 2 pyridine protons + H-9), 11.82 (br s, 1H,
NH). IR 1670, 1728, 3214.
1
Compound 5: Yield: 80%; mp 220–222 ꢁC (DMF). H
NMR 3.78 (s, 3H, OMe), 7.09 (d, 2H, ar, J = 9.2 Hz),
7.25–7.92 (m, 7H, ar), 8.72–8.82 (m, 3H, 2 pyridine pro-
tons + H-9), 11.65 (br s, 1H, NH). IR 1723.
6.1.7. 1,2-Dihydro-4-methanesulfonylamino-2-(4-methoxy-
A
Compound 6: Yield: 97%; mp >300 ꢁC (DMF). 1H
NMR 7.37–7.63 (m, 3H, ar), 7.72–8.06 (m, 6H, ar),
8.78-8.98 (m, 3H, 2 pyridine protons + H-9), 11.77 (s,
1H, NH). IR 1715.
phenyl)-1,2,4-triazolo[4,3-a]quinoxalin-1-one (11).
solution of methanesulfonyl chloride (6.5 mmol,
0.5 mL) in anhydrous dichloromethane (5 mL) was
added at room temperature to a suspension of com-
pound 2316 (0.65 mmol, 0.2 g) in anhydrous dichloro-
methane (30 mL) and anhydrous pyridine (9.1 mmol,
0.71 mL). The mixture was refluxed for 6 h and then a
second portion of methanesulfonyl chloride (9.1 mmol,
0.72 mL) in anhydrous dichloromethane (5 mL) was
added. The suspension was heated at reflux for further
8 h. After cooling at room temperature, the solid was
collected by filtration and washed with water and
washed with water. Yield: 50%; mp 278–280 ꢁC (acetoni-
trile). 1H NMR 3.33 (s, 3H, SO2Me), 3.80 (s, 3H, OMe),
7.11 (d, 2H, ar, J = 8.8 Hz), 7.38–7.43 (m, 2H, ar), 7.78–
7.89 (m, 3H, ar), 8.62–8.67 (m, 1H, H-9), 11.45 (br s,
1H, NH). IR 1157, 1278, 1723, 3190.
6.1.4. General procedure for the synthesis of 4-benze-
nesulfonylamino-1,2-dihydro-2-phenyl-1,2,4-triazolo[4,3-a]-
quinoxalin-1-one (7) and 4-benzenesulfonylamino-1,2-
dihydro-2-(4-methoxyphenyl)-1,2,4-triazolo[4,3-a]quinox-
alin-1-one (8). A mixture of compound 22 or 2316
(1.08 mmol) and benzenesulfonylchloride (14 mmol,
1.78 mL) in anhydrous dichloromethane (30 mL) and
anhydrous pyridine (15 mmol, 1.2 mL) was refluxed un-
til the disappearance of the starting 4-amino derivative
(3–6 days). The mixture was cooled at room temperature
and the solid collected by filtration and washed with
water.
Compound 7: Yield: 64%; mp >300 ꢁC (DMF). 1H
NMR 7.36–7.66 (m, 9H, ar), 7.97 (d, 2H, ar,
J = 6.2 Hz), 8.01–8.05 (m, 2H, ar), 8.62–8.66 (m, 1H,
H-9), 11.4 (br s, 1H, NH). IR 1720, 3300.
6.1.8. 1,2-Dihydro-4-methanesulfonylamino-6-nitro-2-phenyl-
1,2,4-triazolo[4,3-a]quinoxalin-1-one (12). A mixture of
compound 2417 (0.77 mmol) and methanesulfonyl chlo-
ride (58.4 mmol, 4.5 mL) in anhydrous dichloromethane
(30 mL) and anhydrous pyridine (10.5 mmol, 0.83 mL)
was refluxed for 24 h. Then, a second portion of methane-
sulfonyl chloride (58.4 mmol, 4.5 mL) was added to the
mixture and, after refluxing for 24 h, a third portion of
the same amount of methanesulfonyl chloride was added.
The suspension was heated for further 24 h and then
cooled at room temperature. The solid was collected by
filtration and washed with water. Yield: 40%; mp
Compound 8: Yield: 77%; mp >300 ꢁC (DMF). 1H
NMR 3.78 (s, 3H, OMe), 7.07 (d, 2H, ar, J = 9.2 Hz),
7.37–7.42 (m, 2H, ar), 7.44–8.06 (m, 8H, ar), 8.59–8.65
(m, 1H, H-9), 11.5 (br s, 1H, NH). IR 1723, 3266.
6.1.5. 4-Benzenesulfonylamino-1,2-dihydro-6-nitro-2-phe-
nyl-1,2,4-triazolo[4,3-a]quinoxalin-1-one (9). A mixture
of compound 2417 (0.9 mmol) and benzenesulfonylchlo-
ride (2.6 mmol, 0.20 mL) in anhydrous pyridine (5 mL)
was refluxed for 36 h. After cooling at room tempera-
ture, the solid was collected by filtration and washed
with water. Yield: 81%; mp 222–224 ꢁC (DMF). 1H
NMR 7.37–7.64 (m, 7H, ar), 7.96–8.21 (m, 5H, ar),
8.99 (d, 1H, H-9, J = 7.8 Hz), 11.3 (br s, 1H, NH). IR
1378, 1536, 1717, 3125.
1
>300 ꢁC (AcOH). H NMR 3.27 (s, 3H, SO2Me), 7.38
(t, 1H, ar, J = 7.3 Hz), 7.58–7.66 (m, 3H, ar), 8.02 (d,
2H, ar, J = 8.1 Hz) 8.10–8.18 (m, 1H, ar), 9.01 (d, 1H,
H-9, J = 7.7 Hz). IR 1121, 1295, 1376, 1533, 1722, 3110.
6.1.9. 1,2-Dihydro-4-bismethanesulfonylamino-2-phenyl-
1,2,4-triazolo[4,3-a]quinoxalin-1-one (13). A mixture of
compound 2216 (0.75 mmol) and methanesulfonyl chlo-
ride (15.4 mmol, 1.19 mL) in anhydrous dichlorometh-
ane (30 mL) and anhydrous pyridine (10.5 mmol,
0.83 mL) was refluxed for 15 h. After evaporation of
the solvent at reduced pressure, the residue was taken
up with water (10 mL) and the solid collected by filtra-
tion. Yield: 85%; mp 277–278 ꢁC (DMF/EtOH). 1H
NMR 3.88 (s, 6H, 2SO2Me), 7.41 (t, 1H, ar,
J = 7.3 Hz), 7.55–7.71 (m, 3H, ar), 7.86 (t, 1H, ar,
J = 7.3 Hz), 8.02–8.08 (m, 3H, ar), 8.79 (d, 1H, H-9,
J = 8.1 Hz). IR 1730.
6.1.6.
1,2-Dihydro-4-methanesulfonylamino-2-phenyl-
1,2,4-triazolo[4,3-a]quinoxalin-1-one (10). A solution of
methanesulfonyl chloride (2.52 mmol, 0.2 mL) in anhy-
drous pyridine (1 mL) was added at room temperature
to a suspension of compound 2216 (2.52 mmol, 0.7 g)
in anhydrous pyridine (10 mL). The mixture was re-
fluxed for 12 h, then a second portion of methanesulfo-
nyl chloride (5 mmol, 0.4 mL) in anhydrous pyridine
(1 mL) was added and the reflux continued. Further five

6098
V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
6.1.10. 1,2-Dihydro-4-bismethanesulfonylamino-2-(4-
methoxyphenyl)-1,2,4-triazolo[4,3-a]quinoxalin-1-one
(14). The title compound was prepared by heating com-
pound 2316 (0.65 mmol) and methanesulfonyl chloride
(42 mmol, 3.2 mL) in anhydrous dichloromethane
(30 mL) and anhydrous pyridine (10.5 mmol, 0.83 mL)
at reflux for 38 h. Evaporation of the solvent at reduced
pressure afforded a residue which was treated with water
(10 mL). The solid was collected by filtration. Yield:
60%; mp 285–287 ꢁC (EtOAc). H NMR 3.81 (s, 3H,
OMe), 3.86 (s, 6H, 2SO2Me), 7.12 (d, 2H, ar,
J = 9.1 Hz), 7.61–7.94 (m, 4 H, ar), 8.04 (d, 1H, ar,
J = 8.1 Hz), 8.78 (d, 1H, H-9, J = 8.4 Hz). IR 1717.
1H, H-9, J = 8.1 Hz), 11.5 (br s, 1H, NH), 12.5 (br s,
1H, NH). IR 1695, 1730, 3300.
6.1.13.
1,2-Dihydro-4-(3-iodophenylureido)-2-phenyl-
1,2,4-triazolo[4,3-a]quinoxalin-1-one (19). 3-Iodopheny-
lisocyanate (1.3 mmol) was added to a hot solution of
derivative 2216 (0.9 mmol) in anhydrous tetrahydrofu-
ran (10 mL), under nitrogen atmosphere. The mixture
was refluxed for 3 h. After cooling at room temperature,
the solid was collected by filtration and washed with
water and diethyl ether. Yield: 95%; mp 225–226 ꢁC
(AcOH). 1H NMR 7.18–7.65 (m, 8H, ar), 7.82–7.98
(m, 1H, ar), 8.15–8.30 (m, 3H, ar), 8.50–8.65 (m, 1H,
H-9), 11.05 (br s, 1H, NH), 12.01 (br s, 1H, NH). IR
1690, 1730, 3180, 3290.
1
6.1.11. General procedure for the synthesis of 4-benzy-
lureido-1,2-dihydro-2-phenyl-1,2,4-triazolo[4,3-a]quinoxa-
lin-1-one (15), 4-benzylureido-1,2-dihydro-2-(4-methoxyphenyl)-
1,2,4-triazolo[4,3-a]quinoxalin-1-one (16), and 4-benzyl-
ureido-1,2-dihydro-6-nitro-2-phenyl-1,2,4-triazolo[4,3-a]
quinoxalin-1-one (17). Benzylisocyanate (3.25 mmol) was
added to a hot solution of the 4-amino derivative 22,
2316 or 2417 (1.3 mmol) in anhydrous tetrahydrofuran
(10 mL), under nitrogen atmosphere. The mixture was
refluxed for about 3–5 h. After cooling at room tempe-
rature, the solid was collected by filtration and washed
with water and diethyl ether.
6.1.14. General Procedure for the synthesis of 4-benzyl-
oxy-1,2-dihydro-2-aryl-1,2,4-triazolo[4,3-a]quinoxalin-1-
ones (20–21). A suspension of sodium hydride (60% dis-
persion in mineral oil, 2.28 mmol) in anhydrous dimeth-
ylsulfoxide (7 mL) was stirred at 60 ꢁC for 1 h, under
nitrogen atmosphere. After cooling the suspension at
40 ꢁC, benzyl alcohol (2.28 mmol, 0.23 mL) was added.
The temperature of the mixture was left to rise to
20 ꢁC, then a solution of the 4-chloro derivative 2516
or 2616 (1.9 mmol) in anhydrous dimethylsulfoxide
(3 mL) was slowly added, keeping the temperature of
the reaction at 20 ꢁC. After the addition was completed,
the mixture was stirred for about 12–15 h at room tem-
perature, then it was diluted with water (20 mL). The
solution was extracted with chloroform (3· 30 mL).
The organic phases were collected, washed with water
(3· 30 mL), and anhydrified (Na2SO4). Evaporation of
the solvent at reduced pressure gave an oil residue which
when treated with diethyl ether solidified. The solid was
collected by filtration and purified by column chroma-
tography (SiO2, eluting system CHCl3/EtOAc, 10:1).
1
Compound 15: Yield: 55%; mp 258–260 ꢁC (DMF). H
NMR 4.60 (d, 2H, CH2, J = 4.8 Hz), 7.22–7.73 (m, 11H,
ar), 8.18 (d, 2H, ar, J = 8.7 Hz), 8.61 (d, 1H, H-9,
J = 8.2 Hz), 9.72 (br s, 1H, NH), 9.87 (br s, 1H,
NH).
1
Compound 16: Yield: 62%; mp 245–247 ꢁC (DMF). H
NMR 3.79 (s, 3H, OMe), 4.53 (d, 2H, CH2, J = 4.8 Hz),
7.09 (d, 2H, ar, J = 8.8 Hz), 7.20–7.45 (m, 7H, ar), 7.65–
7.66 (m, 1H, ar), 8.02 (d, 2H, ar, J = 8.8 Hz), 8.63 (d,
1H, H-9, J = 8.1 Hz), 9.77 (br s, 1H, NH), 9.86 (br s,
1H, NH). IR 1682, 1718, 3192.
Compound 20: Yield: 35%; mp 168–169 ꢁC (nitrometh-
ane). H NMR 5.64 (s, 2H, CH2), 7.35–7.64 (m, 11H,
1
1
Compound 17: Yield: 85%; mp 256-258ꢁC (DMF). H
ar), 8.01 (d, 2H, ar, J = 7.7 Hz), 8.71 (dd, 1H, H-9,
J = 8.4, 2.6 Hz). IR 1730.
NMR 4.50 (d, 2H, CH2, J = 5.1 Hz), 7.22–7.65 (m,
9H, ar), 8.02 (d, 1H, ar, J = 8.1 Hz), 8.15 (d, 2H, ar,
J = 8.0 Hz), 8.89 (d, 1H, H-9, J = 8.2 Hz), 9.82 (br s,
1H, NH), 10.66 (br s, 1H, NH). IR 1686, 1729,
3184.
Compound 21: Yield: 48%; mp 200–202 ꢁC (AcOH). 1H
NMR 3.89 (s, 3H, OMe), 5.61 (s, 2H, CH2), 7.07 (d, 2H,
ar, J = 7.3 Hz), 7.35–7.88 (m, 10H, ar), 8.67 (d, 1H, H-9,
J = 8.2 Hz). IR 1716.
6.1.12. 4-Benzoylureido-1,2-dihydro-2-phenyl-1,2,4-triaz-
olo[4,3-a]quinoxalin-1-one (18). A mixture of derivative
2216 (1.5 mmol) and benzoylisocyanate (2.32 mmol) in
anhydrous tetrahydrofuran was refluxed for 6 h under
nitrogen atmosphere. The suspension was cooled at
room temperature, the solid was collected by filtration,
and washed with diethyl ether. It was not possible to
purify the crude compound 18 since it resulted unstable
upon heating. In fact, recrystallization from nitrometh-
ane or DMF caused, respectively, its partial and total
degradation to the starting 4-amino derivative. Never-
theless, TLC analysis (CHCl3/MeOH 9.5:0.5), 1H
NMR spectrum, and melting point of crude 18 showed
that it was pure enough to be tested in the binding as-
says. Yield: 82%; mp 242–243 ꢁC (crude). 1H NMR
7.39–7.78 (m, 9H, ar), 8.06–8.09 (m, 4H, ar). 8.75 (d,
6.2. Computational methodologies
All modeling studies were carried out on a 10 CPU
(PIV-3.0 GHZ and AMD64) linux cluster running under
openMosix architecture.56
Homology modeling, energy calculation, and docking
studies were performed using the Molecular Operating
Environment (MOE, version 2006.08) suite.57
All docked structures were fully optimized without
geometry constraints using RHF/AM1 semi-empirical
calculations. Vibrational frequency analysis was used
to characterize the minima stationary points (zero imag-
inary frequencies). The software package MOPAC

V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
6099
(ver.7),58 implemented in MOE suite, was utilized for all
quantum mechanical calculations.
ventional homology modeling strategy that combined
the Boltzmann-weighted randomized modeling proce-
dure adapted from Levitt61 with a specialized algorithm
for the proper handling of insertions and deletions of
any selected extra-atoms during the energy tests and
minimization stages of the modeling procedure. The li-
gand-based option is very useful when one wishes to
build a homology model in the presence of a ligand
docked to the primary template, or other proteins
known to be complexed with the sequence to be mod-
eled. In this specific case, both model building and
refinement take into account the presence of the ligand
in terms of specific steric and chemical features. In order
to generate an initial ensemble of ligand poses, a con-
ventional docking procedure (see next section for de-
tails) with reduced van der Waals radii (equal to 75%)
6.2.1. Homology model of the human A3 AR. Based on
the assumption that GPCRs share similar TM bound-
aries and overall topology, a homology model of the
hA3 receptor was constructed. First, the amino acid se-
quences of TM helices of the A3 receptor were aligned
with those of bovine rhodopsin, guided by the highly
conserved amino acid residues, including the DRY mo-
tif (D3.49, R3.50, and Y3.51) and three proline residues
(P4.60, P6.50, and P7.50) in the TM segments of
GPCRs. The same boundaries were applied for the
TM helices of the A3 receptor as they were identified
from the X-ray crystal structure for the corresponding
sequences of bovine rhodopsin,59 the CR coordinates
of which were used to construct the seven TM helices
for the hA3 receptor. The loop domains of the hA3
receptor were constructed by the loop search method
implemented in MOE. In particular, loops are modeled
first in random order. For each loop, a contact energy
function analyzes the list of candidates collected in the
segment searching stage, taking into account all atoms
already modeled and any atoms specified by the user
as belonging to the model environment. These energies
are then used to make a Boltzmann-weighted choice
from the candidates, the coordinates of which are then
copied to the model. Any missing side-chain atoms are
modeled using the same procedure. Side chains belong-
ing to residues the coordinates of which were copied
from a template are modeled first, followed by side
chains of modeled loops. Outgaps and their side chains
are modeled last. Special caution has to be given to the
second extracellular (EL2) loop, which has been de-
scribed in bovine rhodopsin as folding back over trans-
membrane helices59 and, therefore, limiting the size of
the active site. Hence, amino acids of this loop could
be involved in direct interactions with the ligands. A
driving force to this peculiar fold of the EL2 loop might
be the presence of a disulfide bridge between cysteines in
TM3 and EL2. Since this covalent link is conserved in
all receptors modeled in the current study, the EL2 loop
was modeled using a rhodopsin-like constrained geome-
try around the EL2-TM3 disulfide bridge. After the hea-
vy atoms were modeled, all hydrogen atoms were added,
and the protein coordinates were then minimized with
MOE using the AMBER94 force field.60 The minimiza-
tions were carried out by the 1000 steps of steepest des-
cent followed by conjugate gradient minimization until
˚
and an increased Coulomb-vdW cutoff (cutoff on 10 A;
˚
cutoff on 12 A) was performed. For each pose, a homol-
ogy model is then generated to accommodate the ligand
by reorienting nearby side chains. These residues and the
ligand are then locally minimized. Finally, each ligand is
re-docked into its corresponding low-energy protein
structures and the resulting complexes are ranked
according to MOEScore.57
Different quantitative measurements of molecular vol-
ume of the receptor binding cavities have been carried
out using MOE suite.57 Prediction of antagonist–receptor
complex stability (in terms of corresponding pKi value)
and the quantitative analysis for non-bonded intermolec-
ular interactions (H-bonds, transition metal, water
bridges, and hydrophobic) were calculated and visualized
using several tools implemented into MOE suite.57
6.2.3. Molecular docking of the hA3 AR antagonists. All
antagonist structures were docked into the hypothetical
TM binding site using the MOE-dock tool, part of the
MOE suite. Searching is conducted within a user-speci-
fied 3D docking box (the standard protocol selects all
˚
atoms inside 12 A from the center of mass of the binding
cavity), using the Tabu Search62 protocol (standard
parameters are 1000 steps/run, 10 attempts/step, and
10 Tabu list length), and the MMFF94 force field63
MOE-Dock performs a user-specified number of inde-
pendent docking runs (50 in our specific case) and writes
the resulting conformations and their energies in a
molecular database file. The resulting docked complexes
were subjected to MMFF94 energy minimization until
ꢀ1
˚
A .
the rms of conjugate gradient was <0.1 kcal molꢀ1
the rms gradient of the potential energy was less than
ꢀ1. Protein stereochemistry evaluation
˚
A
was performed by several tools (Ramachandran and
Chi plots measure phi/psi and chi1/chi2 angles, clash
contact reports) implemented in MOE suite.57
Charges for the ligands were imported from the MO-
PAC output files. To better refine all antagonist–recep-
tor complexes, a rotamer exploration of all side chains
involved in the antagonist-binding was carried out. Rot-
amer exploration methodology was implemented in
MOE suite.57
0.1 kcal molꢀ1
6.2.2. Ligand-based homology modeling. We have re-
cently revisited the rhodopsin-based model of the hu-
man A3 receptor in its resting state (antagonist-like
state), taking into account a novel strategy to simulate
the possible receptor reorganization induced by the
antagonist-binding.31 We called this new strategy li-
gand-based homology modeling. Briefly, ligand-based
homology modeling technique is an evolution of a con-
6.3. Pharmacology
6.3.1. Bovine A1 and A2A receptor binding. Displacement
of [3H]DPCPX from A1 ARs in bovine cortical mem-
branes and [3H]CGS 21680 from A2A ARs in bovine
striatal membranes was performed as described in Ref.
64 and Ref. 65, respectively.

6100
V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
6.3.2. Human A1, A2A, and A3 receptor binding. Binding
experiments at hA1 and hA2A ARs, stably expressed in
CHO cells, were performed as previously described in
Ref. 66 using [3H]DPCPX and [3H]NECA, respectively,
as radioligands. Displacement of [125I]AB-MECA from
hA3 AR, stably expressed in CHO cells, was performed
as reported in Ref. 18.
ing the medium and adding 0.4 N HCl. After 30 min the
lysate was neutralized with KOH 4 N and the suspen-
sion was centrifuged at 800g for 5 min. To determine
cyclic AMP production, the binding protein, prepared
from beef adrenal glands, was incubated with [3H]cAMP
(2 nM) in distilled water, 50 lL of cell lysate, or stan-
dard cAMP (0–16 pmol) at 4 ꢁC for 150 min in a total
volume of 300 lL. Bound radioactivity was separated
by rapid filtration through GF/C glass fiber filters and
washed twice with 4 mL of 50 mM Tris/HCl, pH 7.4.
The radioactivity was measured by liquid scintillation
spectrometry.
6.3.3. Rat A1 receptor binding. Rat brain cortex was dis-
sected from male Wistar rats. Membrane preparation
was carried out as previously reported for the bovine
membrane preparation.65 The A1 binding assays were
performed in triplicate by incubating aliquots of brain
cortex membranes (40–50 lg of protein) at 25 ꢁC for
180 min in 0.5 mL of binding buffer (50 mM Tris/HCl,
2 mM MgCl2, pH 7.7) containing 0.2–0.5 nM
[3H]DPCPX. Non-specific binding was defined in the
presence of 20 lM R-PIA. Incubation was terminated
by rapid filtration through Whatman GF/C glass micro-
fiber filters and washing twice with 4 mL of ice-cold buf-
fer. KD value on brain cortex was 0.51 nM.67
6.3.6. Data analysis. The concentration of the tested
compounds that produced 50% inhibition of specific
[3H]DPCPX, [3H]NECA, [3H]CGS 21680 and [125I]AB-
MECA binding (IC50) was calculated using a non-linear
regression method implemented by the InPlot program
(Graph-Pad, San Diego, CA, USA) with five concentra-
tions of displacer, each performed in triplicate. Inhibi-
tion constants (Ki) were calculated according to the
Cheng–Prusoff equation.70 The dissociation constant
(Kd) values of [3H] DPCPX and [3H]CGS 21680 in cor-
tical and striatal bovine brain membranes were 0.3 nM
and 14 nM, respectively.The Kd values of [3H]DPCPX,
[3H]NECA and [125I]AB-MECA in hA1, hA2A and
hA3 ARs in CHO cell membranes were 3 nM, 30 nM,
and 1.4 nM, respectively. EC50 values obtained in cAMP
assays were calculated by non-linear regression analysis
using the equation for a sigmoid concentration–response
curve (Graph-Pad, San Diego, CA, USA).
6.3.4. Rat A3 receptor binding. Binding of [3H] (R)-PIA
(37 Ci/mmol) to rat testis membranes was measured in
the presence of DPCPX (150 nM) as previously de-
scribed.23,68 Briefly, fresh testicular tissue from Wistar
rats was dissected free of epididymis and membranes
were prepared as described.68 Rat testis membranes
(0.1–0.2 mg of protein) and [3H] (R)-PIA 4 nM were
incubated in 0.5 mL total volume of 50 mM Tris/
HCl (pH 7.4), 1 mM EDTA, and 10 mM MgCl2 buffer
in the presence of 150 nM DPCPX to block A1 aden-
osine receptors. Non-specific binding was determined
in the presence of 15 lM (R)-PIA. Binding reactions
were terminated by filtration through Whatman GF/
filters under reduced pressure. Filters were washed
three times with 5 mL of ice-cold buffer and intro-
duced into scintillation vials. The radioactivity was
counted in 4 mL of scintillation cocktail in a scintilla-
tion counter.
6.3.7. Electrophysiological assays: slice preparation. All
animal procedures were carried out according to the
European Community Guidelines for Animal Care,
DL 116/92, application of the European Communities
Council Directive (86/609/EEC). Experiments were car-
ried out on rat hippocampal slices, prepared as previ-
ously described.71 Male Wistar rats (Harlan Italy;
Udine Italy, 150–200 g body weight) were killed with a
guillotine under anesthesia with ether and their hippo-
campi were rapidly removed and placed in ice-cold oxy-
genated (95% O2–5% CO2) aCSF of the following
composition (mM): NaCl 124, KCl 3.33, KH2PO4
1.25, MgSO4 1.4, CaCl2 2.5, NaHCO3 25, and D-glucose
10. Slices (400 lm thick) were cut using a McIlwain tis-
sue chopper (The Mickle Lab. Engineering, Co. Ltd,
Gomshall, UK) and kept in oxygenated aCSF for at
least 1 h at room temperature. A single slice was then
placed on a nylon mesh, completely submerged in a
small chamber (0.8 mL), and superfused with oxygen-
ated aCSF (30–32 ꢁC) at a constant flow rate of 1.5/
2 mL minꢀ1. The treated solutions reached the prepara-
tion in 90 s and this delay was taken into account in our
calculations.
6.3.5. Measurement of cAMP Levels on CHO cells
transfected with human A2B and A3 ARs. Intracellular
cAMP levels were measured using a competitive protein
binding method.69 CHO cells (ꢁ60,000), stably express-
ing hA2B or hA3 ARs, were plated in 24-well plates.
After 48 h, the medium was removed, and the cells were
incubated at 37 ꢁC for 15 min with 0.5 mL di DMEM in
the presence of Ro 20-1724 (4-(3-butoxy-4-methoxyben-
zyl)imidazolidin-2-one) (20 lM) and adenosine deami-
nase (1U/mL). A stock 1 mM solution of the tested
compound was prepared in DMSO, and subsequent
dilutions were accomplished in distilled water. The
antagonistic profile of the new compound toward
hA2B AR was evaluated assessing its ability to inhibit
100 nM NECA-mediated accumulation of cAMP. The
antagonistic profile of the new compound toward hA3
AR was evaluated by assessing its ability to counteract
100 nM NECA-mediated inhibition of cAMP accumula-
tion stimulated by 1 lM forskolin. Cells were incubated
in the reaction medium (15 min at 37 ꢁC) with different
compound concentrations (1 nM–10 lM) and then trea-
ted with NECA. The reaction was terminated by remov-
6.3.7.1. Extracellular recording. Test pulses (80 ls,
0.066 Hz) were delivered through a bipolar nichrome
electrode positioned in the stratum radiatum. Evoked
extracellular potentials were recorded with glass micro-
electrodes (2–10 MX, Clark Electromedical Instruments,
Panghourne, UK) filled with 150 mM NaCl, placed in
the CA1 region of the stratum radiatum. Responses

V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
6101
were amplified (BM 622, Mangoni, Pisa, Italy), digitized
(sample rate, 33.33 kHz), and stored for later analysis
using LTP (version 2.30D) software facilities.72 Stimu-
lus–response curves were obtained by gradual increases
in stimulus strength at the beginning of each experiment,
until a stable baseline of evoked response was reached.
The test stimulus pulse was then adjusted to produce a
fEPSP whose slope and amplitude were 40–50% of the
maximum and were kept constant throughout the exper-
iment. The fEPSP amplitude was routinely measured
and expressed as the percentage of the average ampli-
tude of the potentials measured during the 5-min pre-
ceding exposure of the hippocampal slices to OGD.
to Chemical Computing Group for the scientific and
technical partnership.
References and notes
1. Fredholm, B. B.; IJzerman, A. P.; Jacobson, K. A.; Klotz,
K. N.; Linden, J. Pharmacol. Rev. 2001, 53, 527.
2. Jacobson, K. A.; Knutsen, L. J. S. In Purinergic and
Pyrimidinergic Signalling; Abbracchio, M. P., Williams,
M., Eds.; Handbook of Experimental Pharmacology:
Berlin, 2001; Vol. 151/1, p 129.
3. Abbracchio, M. P.; Brambilla, R.; Kim, H. O.; von
Lubitz, D. K. J. E.; Jacobson, K. A.; Cattabeni, F. Mol.
Pharmacol. 1995, 48, 1083.
In all hippocampal slices, simultaneously with fEPSP
amplitude, we also recorded AD in d.c. (direct current)
mode and measured from the beginning of OGD as neg-
ative d.c. shifts.
4. Ali, H.; Choi, O. H.; Fraundorfer, P. F.; Yamada, K.;
Gonzaga, H. M. S.; Beaven, M. A. J. Pharmacol. Exp.
Ther. 1996, 276, 837.
5. Shneyvays, V.; Leshem, D.; Zinman, T.; Mamedova, L.
K.; Jacobson, K. A.; Shainberg, A. Am. J. Physiol. Heart
Circ. Physiol. 2005, 288, H2792.
6.3.7.2. Application of drugs and OGD. OGD was ob-
tained by perfusing the slice for 7 min with aCSF with-
out glucose and gassed with nitrogen (95% N2–5%
CO2).73 When slices were subjected to 7 min of OGD,
if the recovery of fEPSP amplitude after 15 min of reper-
fusion with glucose-containing and normally oxygen-
ated aCSF was 615% of the pre-ischemic value, a
second slice from the same rat was submitted to a 7-
min OGD insult in the presence of compound 13. To
confirm the result obtained in the treated group, a third
slice was taken from the same rat and another 7-min
OGD was performed in control conditions, to verify
that no difference between slices was caused by the time
gap between the experiments.
6. Linden, J. Trends Pharmacol. Sci. 1994, 15, 298.
7. Liang, B. T.; Jacobson, K. A. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 6995.
8. Young, H. W.; Molina, J. G.; Dimina, D.; Zhong, H.;
Jacobson, M.; Chan, L. N.; Chan, L.-N. L.; Chan, T.-S.;
Lee, J. J.; Blackburn, M. R. J. Immunol. 2004, 173, 1380.
9. Merighi, S.; Mirandola, P.; Varani, K.; Gessi, S.; Leung,
E.; Baraldi, P. G.; Tabrizi, M. A.; Borea, P. A. Pharmacol.
Ther. 2003, 100, 31.
10. Jacobson, K. A.; Gao, Z.-G. Nat. Rev. Drug Disc. 2006, 5,
247.
11. Lee, H. T.; Ota-Setlik, A.; Xu, H.; D’Agati, V. D.;
Jacobson, M. A.; Emala, C. W. Am. J. Physiol. Renal
Physiol. 2003, 284, 267.
12. Brambilla, R.; Cattabeni, F.; Ceruti, S.; Barbieri, D.;
Franceschi, C.; Kim, Y.-C.; Jacobson, K. A.; Klotz, K.-
N.; Lohse, M. J.; Abbracchio, M. P. Naunyn-Schmiede-
berg’s Arch. Pharmacol. 2000, 361, 225.
13. Pugliese, A. M.; Coppi, E.; Spalluto, G.; Corradetti, R.;
Pedata, F. Br. J. Pharmacol. 2006, 147, 524.
14. Merighi, S.; Benini, A.; Mirandola, P.; Gessi, S.; Varani,
K.; Leung, E.; Maclennan, S.; Borea, P. A. Biochem.
Pharmacol. 2006, 72, 19.
15. Colotta, V.; Catarzi, D.; Varano, F.; Cecchi, L.; Filacchi-
oni, G.; Martini, C.; Trincavelli, L.; Lucacchini, A. J.
Med. Chem. 2000, 43, 3118.
The selective A3 adenosine receptor antagonist was ap-
plied 15 min before, during, and 5 min after OGD. Con-
centration of the selective adenosine A3 receptor
antagonist was chosen on the basis of Ki values on
hA3 AR.
6.3.7.3. Statistical analysis. Data were analyzed using
Prism 3.02 software (Graphpad Software, San Diego,
CA, USA). All numerical data are expressed as mean-
s
SEM. Data were tested for statistical significance
16. Colotta, V.; Catarzi, D.; Varano, F.; Cecchi, L.; Filacchi-
oni, G.; Martini, C.; Trincavelli, L.; Lucacchini, A. J.
Med. Chem. 2000, 43, 1158.
by unpaired two-tailed Student’s t test. A value of
P < 0.05 was considered significant.
17. Colotta, V.; Catarzi, D.; Varano, F.; Filacchioni, G.;
Martini, C.; Trincavelli, L.; Lucacchini, A. Bioorg. Med.
Chem. 2003, 11, 3541.
Acknowledgments
18. Colotta, V.; Catarzi, D.; Varano, F.; Filacchioni, G.;
Martini, C.; Trincavelli, L.; Lucacchini, A. Bioorg. Med.
Chem. 2003, 11, 5509.
19. Colotta, V.; Catarzi, D.; Varano, F.; Calabri, F. R.; Lenzi,
O.; Filacchioni, G.; Trincavelli, L.; Martini, C.; Deflorian,
F.; Moro, S. J. Med. Chem. 2004, 47, 3580.
20. Catarzi, D.; Colotta, V.; Varano, F.; Calabri, F. R.; Lenzi,
O.; Filacchioni, G.; Trincavelli, L.; Martini, C.; Tralli, A.;
Montopoli, C.; Moro, S. Bioorg. Med. Chem. 2005, 13,
705.
21. Catarzi, D.; Colotta, V.; Varano, F.; Lenzi, O.; Filacchi-
oni, G.; Trincavelli, L.; Martini, C.; Montopoli, C.; Moro,
S. J. Med. Chem. 2005, 48, 7932.
This work was supported by a grant of the Italian Min-
istry for University and Research (MIUR, FIRB
RBNE03YA3L project). The molecular modeling work
coordinated by S.M. was carried out with financial sup-
port from the University of Padova, Italy, and the Ital-
ian Ministry for University and Research (MIUR),
Rome, Italy. The electrophysiological work was sup-
ported by a grant from the Italian Ministry for Univer-
sity and Research (MIUR) and by Ente Cassa di
Risparmio di Firenze (2004/0747). We thank Dr. Karl-
Norbert Klotz of the University of Wuzburg, Germany,
¨
for providing cloned hA1, hA2A, hA2B and hA3 recep-
tors expressed in CHO cells. S.M. is also very grateful
22. Lenzi, O.; Colotta, V.; Catarzi, D.; Varano, F.; Filacchi-
oni, G.; Martini, C.; Trincavelli, L.; Ciampi, O.; Mari-

6102
V. Colotta et al. / Bioorg. Med. Chem. 16 (2008) 6086–6102
ghetti, F.; Morizzo, E.; Moro, S. J. Med. Chem. 2006, 49,
3916.
49. Tanaka, E.; Yamamoto, S.; Kudo, Y.; Mihara, S.;
Higashi, H. J. Neurophysiol. 1997, 78, 891.
23. Colotta, V.; Catarzi, D.; Varano, F.; Capelli, F.; Lenzi, O.;
Filacchioni, G.; Martini, C.; Trincavelli, L.; Ciampi, O.;
Pugliese, A. M.; Pedata, F.; Schiesaro, A.; Morizzo, E.;
Moro, S. J. Med. Chem. 2007, 50, 40.
24. Chang, L. C. W.; Brussee, J.; IJzerman, A. P. Chem.
Biodivers. 2004, 1, 1591.
25. Ukena, D.; Jacobson, K. A.; Padgett, W. L.; Ayala, C.;
Shamim, M. T.; Kirk, K. L.; Olsson, R. O.; Daly, J. W.
FEBS Lett. 1986, 209, 122.
26. Muller, C. Exp. Opin. Ther. Patents 1997, 7, 419.
27. Moro, S.; Deflorian, F.; Spalluto, G.; Pastorin, G.;
Cacciari, B.; Kim, S.-K.; Jacobson, K. A. Chem. Commun.
(Camb.) 2003, 24, 2949.
50. Yamamoto, S.; Tanaka, E.; Shoji, Y.; Kudo, Y.; Inokuchi,
H.; Higashi, H. J. Neurophysiol. 1997, 78, 903.
51. Nieber, K.; Hentschel, S. In Proceedings of the 8th
International Symposium on Adenosine and Adenine Nucle-
otides, Ferrara, Italy, May 24–28, 2006.
52. Macek, T. A.; Schaffhauser, H.; Conn, P. J. J. Neurosci.
1998, 18, 6138.
53. Yli-Kauhaluoma, J. T.; Ashkey, J. A.; Lo, C.-H.; Tuker,
L.; Wolfe, M. M.; Janda, K. D. J. Am. Chem. Soc. 1995,
117, 7041.
54. Yu, D.-W.; Gatley, S. J.; Wolf, A. P.; MacGregor, R. R.;
Dewey, S. L.; Fowler, J. S.; Schlyer, D. J. J. Med. Chem.
1992, 35, 2178.
28. Moro, S.; Spalluto, G.; Jacobson, K. A. Trends Pharma-
col. Sci. 2005, 26, 44.
55. Schnetteler, R. A.; Dage, R. C.; Palopoli, F. P. J. Med.
Chem. 1986, 29, 860.
29. Moro, S.; Deflorian, F.; Bacilieri, M.; Spalluto, G. Curr.
Med. Chem. 2006, 13, 639.
30. Moro, S.; Bacilieri, M.; Deflorian, F.; Spalluto, G. New J.
Chem. 2006, 30, 301.
31. Moro, S.; Deflorian, F.; Bacilieri, M.; Spalluto, G. Curr.
Pharm. Des. 2006, 12, 2175.
32. Tafi, A.; Bernardini, C.; Botta, M.; Corelli, F.; Andreini,
M.; Martinelli, A.; Ortore, G.; Baraldi, P. G.; Fruttarolo,
F.; Borea, P. A.; Tuccinardi, T. J. Med. Chem. 2006, 49,
4085.
33. Kim, S. K.; Gao, Z. G.; Jeong, L. S.; Jacobson, K. A. J.
Mol. Graphics Modell. 2006, 25, 562.
34. Jiang, Q.; Lee, B. X.; Glashofer, M.; van Rhee,
A. M.; Jacobson, K. A. J. Med. Chem. 1997, 40,
2588.
35. Jiang, Q.; Van Rhee, A. M.; Kim, J.; Yehle, S.; Wess, J.,
et al. Mol. Pharmacol. 1996, 50, 512.
36. Latini, S.; Pedata, F. J. Neurochem. 2001, 79, 463.
37. Pugliese, A. M.; Coppi, E.; Volpini, R.; Cristalli, G.;
Corradetti, R.; Jeong, L. S.; Jacobson, K. A.; Pedata, F.
Biochem. Pharmacol. 2007, 74, 768.
56. OpenMosix: http://www.openMosix.org, 2004.
57. MOE (The Molecular Operating Environment) Version
2006.08, software available from Chemical Computing
Group Inc., 1010 Sherbrooke Street West, Suite 910,
Montreal, Canada H3A 2R7. http://www.chemcomp.com.
58. MOPAC 7, J.J.P. Stewart, Fujitsu Limited, Tokyo, Japan,
1993.
59. Palczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.;
Motoshima, H., et al. Science 2000, 289, 739.
60. Cornell, W. D. C. P.; Bayly, C. I.; Gould, I. R.; Merz, K.
M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.;
Caldwell, J. W.; Kollman, P. A. J. Am. Chem. Soc.
1995, 117, 5179.
61. Levitt, M. J. Mol. Biol. 1992, 226, 507.
62. Baxter, C. A.; Murray, C. W.; Clark, D. E.; Westhead, D.
R.; Eldridge, M. D. Funct. Genet. 1998, 33, 367.
63. Halgren, T. J. Comput. Chem. 1996, 17, 490.
64. Manetti, F.; Schenone, S.; Bondavalli, F.; Brullo, C.;
Bruno, O.; Ranise, A.; Mosti, L.; Menozzi, G.; Fossa, P.;
Trincavelli, M. L.; Martini, C.; Martinelli, A.; Tintori, C.;
Botta, M. J. Med. Chem. 2005, 48, 7172.
38. Obeidat, A. S.; Andrew, R. D. Eur. J. Neurosci. 1998, 10,
3451.
39. Somjen, G. G. Physiol. Rev. 2001, 81, 1065.
40. Latini, S.; Bordoni, F.; Pedata, F.; Corradetti, R. Br. J.
Pharmacol. 1999, 127, 729.
41. Pearson, T.; Damian, K.; Lynas, R. E.; Frenguelli, B. G.
J. Neurochem. 2006, 97, 1357.
43. Jeong, L. S.; Choe, S. A.; Gunaga, P.; Kim, H. O.; Lee, H.
W.; Lee, S. K.; Tosh, D.; Patel, A.; Palaniappan, K. K.;
Gao, Z. G.; Jacobson, K. A.; Moon, H. R. J. Med. Chem.
2007, 50, 3159.
44. Li, A.-H.; Moro, S.; Melman, N.; Ji, X.-d.; Jacobson, K.
A. J. Med. Chem. 1998, 41, 3186.
45. Kim, Y.-C.; Ji, X.-d.; Jacobson, K. A. J. Med. Chem.
1996, 39, 4142.
46. Maconi, A.; Pastorin, G.; da Ros, T.; Spalluto, G.; Gao,
Z.-G.; Jacobson, K. A.; Baraldi, P. G.; Cacciari, B.;
Varani, K.; Moro, S.; Borea, P. A. J. Med. Chem. 2002,
45, 3579.
47. Ji, X. D.; Gallo-Rodrighuez, C.; Jacobson, K. A. Biochem.
Biophys. Res. Commun. 1994, 203, 570.
48. von Lubitz, D. K. Eu. J. Pharmacol. 1999, 263, 59.
65. Colotta, V.; Catarzi, D.; Varano, F.; Melani, F.; Filac-
chioni, G.; Cecchi, L.; Trincavelli, L.; Martini, C.;
Lucacchini, A. Farmaco 1998, 53, 189.
66. Novellino, E.; Cosimelli, B.; Ehlardo, M.; Greco, G.;
Iadanza, M.; Lavecchia, A.; Rimoli, M. G.; Sala, A.; Da
Settimo, A.; Primofiore, G.; Da Settimo, F.; Taliani, S.;
La Motta, C.; Klotz, K.-N.; Tuscano, D.; Trincavelli, M.
L.; Martini, C. J. Med. Chem. 2005, 48, 8253.
67. Klotz, K.-N.; Vogt, H.; Tawfik-Schlieper, H. Naunyn-
Schmiedeberg’s Arch Pharmacol. 1991, 343, 196.
68. Mazzoni, M. R.; Buffoni, R. S.; Giusti, L.; Lucacchini, A.
J. Recept. Signal Transduct. Res. 1995, 15, 905.
69. Nordstedt, C.; Fredholm, B. Anal. Biochem. 1990, 189,
231.
70. Cheng, Y. C.; Prusoff, W. H. Biochem. Pharmacol. 1973,
22, 3099.
71. Pugliese, A. M.; Latini, S.; Corradetti, R.; Pedata, F. Br.
J. Pharmacol. 2003, 140, 305.
72. Anderson, W. W.; Collingridge, G. L. J. Neurosci.
Methods 2001, 108, 7.
73. Pedata, F.; Latini, S.; Pugliese, A. M.; Pepeu, G. J.
Neurochem. 1993, 61, 284.